未验证 提交 651bee4a 编写于 作者: J juncaipeng 提交者: GitHub

Update support_operation_list (#2104)


* update support_operation_list.md
上级 ee2dddec
...@@ -4,12 +4,18 @@ title: 支持OP列表 ...@@ -4,12 +4,18 @@ title: 支持OP列表
--- ---
## Ops ## Ops
- argmax
- affine_channel
- anchor_generator
- arg_max
- assign
- assign_value
- axpy - axpy
- batch_norm - batch_norm
- beam_search - beam_search
- beam_search_decode - beam_search_decode
- bilinear_interp - bilinear_interp
- box_clip
- box_coder - box_coder
- calib - calib
- calib_once - calib_once
...@@ -23,23 +29,36 @@ title: 支持OP列表 ...@@ -23,23 +29,36 @@ title: 支持OP列表
- depthwise_conv2d - depthwise_conv2d
- dropout - dropout
- elementwise_add - elementwise_add
- elementwise_div
- elementwise_max - elementwise_max
- elementwise_mul - elementwise_mul
- elementwise_sub - elementwise_sub
- equal
- exp - exp
- expand
- fake_dequantize_max_abs - fake_dequantize_max_abs
- fake_quantize_moving_average_abs_max - fake_quantize_moving_average_abs_max
- fake_quantize_range_abs_max
- fc - fc
- feed - feed
- fetch - fetch
- fill_constant - fill_constant
- fill_constant_batch_size_like
- flatten
- flatten2
- floor
- fusion_elementwise_add_activation - fusion_elementwise_add_activation
- fusion_elementwise_div_activation
- fusion_elementwise_max_activation - fusion_elementwise_max_activation
- fusion_elementwise_mul_activation - fusion_elementwise_mul_activation
- fusion_elementwise_sub_activation - fusion_elementwise_sub_activation
- generate_proposals
- graph_op - graph_op
- greater_equal
- greater_than
- gru - gru
- gru_unit - gru_unit
- hard_sigmoid
- im2sequence - im2sequence
- increment - increment
- io_copy - io_copy
...@@ -48,31 +67,42 @@ title: 支持OP列表 ...@@ -48,31 +67,42 @@ title: 支持OP列表
- layout - layout
- layout_once - layout_once
- leaky_relu - leaky_relu
- less_equal
- less_than - less_than
- lod_reset - lod_reset
- log - log
- logical_and - logical_and
- logical_not
- logical_or
- logical_xor - logical_xor
- lookup_table - lookup_table
- lrn - lrn
- matmul
- mean
- mul - mul
- multiclass_nms - multiclass_nms
- nearest_interp - nearest_interp
- negative - negative
- norm - norm
- notequal
- pad2d - pad2d
- pool2d - pool2d
- power - power
- prelu - prelu
- prior_box - prior_box
- range
- read_from_array - read_from_array
- reduce_max - reduce_max
- reduce_mean
- relu - relu
- relu6
- relu_clipped - relu_clipped
- reshape - reshape
- reshape2 - reshape2
- roi_align
- scale - scale
- sequence_expand - sequence_expand
- sequence_expand_as
- sequence_pool - sequence_pool
- sequence_softmax - sequence_softmax
- shape - shape
...@@ -82,27 +112,47 @@ title: 支持OP列表 ...@@ -82,27 +112,47 @@ title: 支持OP列表
- softmax - softmax
- split - split
- square - square
- squeeze
- squeeze2
- stack
- swish - swish
- tanh - tanh
- top_k - top_k
- transpose - transpose
- transpose2 - transpose2
- uniform_random
- unsqueeze
- unsqueeze2
- while - while
- write_to_array - write_to_array
- yolo_box - yolo_box
## Kernels ## Kernels
### Host kernels
- feed
- fetch
- flatten
- flatten2
- multiclass_nms
- reshape
- reshape2
### ARM kernels ### ARM kernels
- argmax
- affine_channel
- anchor_generator
- arg_max
- assign
- assign_value
- axpy - axpy
- batch_norm - batch_norm
- beam_search - beam_search
- beam_search_decode - beam_search_decode
- bilinear_interp - bilinear_interp
- box_clip
- box_coder - box_coder
- calib
- calib_once
- cast - cast
- concat - concat
- conv2d - conv2d
...@@ -113,41 +163,60 @@ title: 支持OP列表 ...@@ -113,41 +163,60 @@ title: 支持OP列表
- depthwise_conv2d - depthwise_conv2d
- dropout - dropout
- elementwise_add - elementwise_add
- elementwise_div
- elementwise_max - elementwise_max
- elementwise_mul - elementwise_mul
- elementwise_sub
- equal
- exp - exp
- expand
- fc - fc
- fill_constant - fill_constant
- floor
- fusion_elementwise_add_activation - fusion_elementwise_add_activation
- fusion_elementwise_div_activation
- fusion_elementwise_max_activation - fusion_elementwise_max_activation
- fusion_elementwise_mul_activation - fusion_elementwise_mul_activation
- fusion_elementwise_sub_activation
- generate_proposals
- greater_equal
- greater_than
- gru - gru
- gru_unit - gru_unit
- hard_sigmoid
- im2sequence - im2sequence
- increment - increment
- is_empty - is_empty
- leaky_relu - leaky_relu
- less_equal
- less_than - less_than
- lod_reset - lod_reset
- log - log
- logical_and - logical_and
- logical_not
- logical_or
- logical_xor - logical_xor
- lookup_table - lookup_table
- lrn - lrn
- matmul
- mul - mul
- multiclass_nms
- nearest_interp - nearest_interp
- negative - negative
- norm - norm
- not_equal
- pad2d - pad2d
- pool2d - pool2d
- power - power
- prelu - prelu
- prior_box - prior_box
- range
- read_from_array - read_from_array
- reduce_max - reduce_max
- reduce_mean
- relu - relu
- relu6
- relu_clipped - relu_clipped
- roi_align
- scale - scale
- sequence_expand - sequence_expand
- sequence_pool - sequence_pool
...@@ -158,32 +227,40 @@ title: 支持OP列表 ...@@ -158,32 +227,40 @@ title: 支持OP列表
- slice - slice
- softmax - softmax
- split - split
- squeeze
- squeeze2
- stack
- swish - swish
- tanh - tanh
- top_k - top_k
- transpose - transpose
- transpose2 - transpose2
- unsqueeze
- unsqueeze2
- while - while
- write_to_array - write_to_array
- yolo_box - yolo_box
### X86 kernels ### X86 kernels
- batch_norm
- concat - concat
- conv2d
- depthwise_conv2d
- dropout
- elementwise_add - elementwise_add
- elementwise_sub - elementwise_sub
- fc - fill_constant_batch_size_like
- fill_constant - gru
- matmul
- mul - mul
- pool2d
- relu - relu
- reshape
- reshape2
- scale - scale
- sequence_expand_as
- sequence_pool
- shape
- slice
- softmax - softmax
- square - square
- squeeze
- squeeze2
### OpenCL kernels ### OpenCL kernels
- conv2d - conv2d
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册